During its CES 2014 press conference, Nvidia has unveiled its new Tegra K1 mobile chip, which packs no less than 192-core Kepler-based GPU.

Previously known under the Logan codename, the Tegra K1 will apparently be available in two versions, as a dual-core Denver-based 64-bit part as well as A15-based 32-bit quad-core part. It is still not clear if both parts will have the same GPU solution, but the one that Nvidia talked earlier today included DirectX 11 support, 192-core GPU with support for Unreal Engine 4 and Open GL 4.4.

According to known and revealed details, the 32-bit ARM A15 based quad-core Tegra K1 version will be clocked at 2.3GHz while Denver-based dual-core 64-bit version will work at 2.5GHz. Nvidia's choice of pushing a dual-core part out is definitely interesting as today, when most companies are pushing for more cores, focus for Nvidia will most likely switch to lower number of higher-performance cores.

Ditching the old Tegra naming scheme, which would place the Tegra 5 name next to Logan, Jen-Hsun noted that it simply did not feel appropriate and that "K" stands for Kepler architecture used for the GPU. Unfortunately, Tegar K1 still lacks native on-chip LTE support and that is definitely a big drawback for Tegra K1.

In any case, Tegra K1 sounds like a great chip with impressive graphics performance but competition is definitely fierce these days and, hopefully, Tegra K1 will do much better than Tegra 4.





Source: Nvidia.com.